SMART: Coordinated Double-Sided Seal Bid Multiunit First Price Auction Mechanism for Cloud-Based TVWS Secondary Spectrum Market

Abstract

Spectrum trading is an important aspect of television white space (TVWS) and it is driven by the failure of spectrum sensing techniques. In spectrum trading, the primary users lease their unoccupied spectrum to the secondary users for a market fee. Although spectrum trading is considered as a reliable approach, it is confronted with a spectrum transaction completion time problem, which negatively impacts on end-users Quality of Service and Quality of Experience metrics. Spectrum transaction completion time is the duration to successfully conduct TVWS spectrum trading. To address this issue, this paper proposes simple mechanism auction reward truthful (SMART), a fast and iterative machine learning-assisted spectrum trading model to address this issue. Simulated results indicate that SMART out-performs referenced VERUM algorithm in three key performance indicators: bit-error rate, instantaneous throughput, and probability of dropped packets by 10%, 5%, and 15%, respectively.

Keywords

• Iterative auction
• machine learning
• game theory
• truthfulness
• mechanism design
• factorial punishment